RepoMirrors/ffmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2024-12-30 03:12:08 +00:00
Code Issues Projects Releases Wiki Activity
116,965 Commits 37 Branches 400 Tags 246 MiB
cfbf10412f
Commit Graph

2606 Commits

Author SHA1 Message Date
Ramiro Polla
c0666d8bed swscale/aarch64/rgb2rgb: add neon implementation for rgb24toyv12 
A55               A76
rgb24toyv12_16_200_c:     36890.6           17275.5
rgb24toyv12_16_200_neon:  12460.1 ( 2.96x)   5360.8 ( 3.22x)
rgb24toyv12_128_60_c:     83205.1           39884.8
rgb24toyv12_128_60_neon:  27468.4 ( 3.03x)  13552.5 ( 2.94x)
rgb24toyv12_512_16_c:     88111.6           42346.8
rgb24toyv12_512_16_neon:  29126.6 ( 3.03x)  14411.2 ( 2.94x)
rgb24toyv12_1920_4_c:     82068.1           39620.0
rgb24toyv12_1920_4_neon:  27011.6 ( 3.04x)  13492.2 ( 2.94x)
 2024-09-06 23:11:13 +02:00
Ramiro Polla
caaec2ea95 swscale/x86/rgb2rgb: disable rgb24toyv12_mmxext for x86_64 
The mmxext implementation is slower than the C version in x86_64.

                                m32               m64
rgb24toyv12_16_200_c:       24942.7           14812.6
rgb24toyv12_16_200_mmxext:  17857.2 ( 1.40x)  17400.4 ( 0.85x)
rgb24toyv12_128_60_c:       56892.9           35616.9
rgb24toyv12_128_60_mmxext:  40730.9 ( 1.40x)  39610.4 ( 0.90x)
rgb24toyv12_512_16_c:       58402.7           37209.4
rgb24toyv12_512_16_mmxext:  44842.4 ( 1.30x)  41136.2 ( 0.90x)
rgb24toyv12_1920_4_c:       54827.4           34737.4
rgb24toyv12_1920_4_mmxext:  51169.9 ( 1.07x)  34818.9 ( 1.00x)
 2024-09-06 23:06:38 +02:00
Ramiro Polla
3604b2403c swscale/rgb2rgb: improve chroma conversion in ff_rgb24toyv12_c 
The current code subsamples by dropping 3/4 pixels to calculate the
chroma components. This commit calculates the average of 4 rgb pixels
before calculating the chroma components, putting it in line with the
mmxext implementation.
 2024-09-06 23:06:32 +02:00
Ramiro Polla
d8848325a6 swscale/aarch64/rgb2rgb: add deinterleaveBytes neon implementation 
A55               A76
deinterleave_bytes_c:             70342.0           34497.5
deinterleave_bytes_neon:          21594.5 ( 3.26x)   5535.2 ( 6.23x)
deinterleave_bytes_aligned_c:     71340.8           34651.2
deinterleave_bytes_aligned_neon:   8616.8 ( 8.28x)   3996.2 ( 8.67x)
 2024-09-06 23:05:09 +02:00
Ramiro Polla
4c824ad391 swscale/x86/rgb2rgb: fix deinterleaveBytes writing past the end of the buffers 2024-09-06 23:05:04 +02:00
Ramiro Polla
f17a6bd200 swscale/x86/rgb2rgb: fix deinterleaveBytes for unaligned dst pointers 2024-09-06 23:05:01 +02:00
Rémi Denis-Courmont
27d28b68da swscale/rgb2rgb: enable R-V V deinterleaveBytes 
T-Head C908:
deinterleave_bytes_c:                               100328.3 ( 1.00x)
deinterleave_bytes_rvv_i32:                          19331.3 ( 5.19x)
deinterleave_bytes_aligned_c:                       100337.5 ( 1.00x)
deinterleave_bytes_aligned_rvv_i32:                  15748.0 ( 6.37x)

SpacemiT X60:
deinterleave_bytes_c:                                95230.6 ( 1.00x)
deinterleave_bytes_rvv_i32:                           9790.3 ( 9.73x)
deinterleave_bytes_aligned_c:                        96564.1 ( 1.00x)
deinterleave_bytes_aligned_rvv_i32:                   7780.1 (12.41x)
 2024-09-04 22:04:11 +03:00
Ramiro Polla
420d443600 swscale/aarch64: cosmetics fix (spaces inside curly braces) 2024-08-26 11:07:49 +02:00
Ramiro Polla
52887683e9 swscale/aarch64: add nv24/nv42 to yuv420p unscaled converter 
A55               A76
nv24_yuv420p_128_c:       4956.1            1267.0
nv24_yuv420p_128_neon:    3109.1 ( 1.59x)    640.0 ( 1.98x)
nv24_yuv420p_1920_c:     35728.4           11736.2
nv24_yuv420p_1920_neon:   8011.1 ( 4.46x)   2436.0 ( 4.82x)
nv42_yuv420p_128_c:       4956.4            1270.5
nv42_yuv420p_128_neon:    3074.6 ( 1.61x)    639.5 ( 1.99x)
nv42_yuv420p_1920_c:     35685.9           11732.5
nv42_yuv420p_1920_neon:   7995.1 ( 4.46x)   2437.2 ( 4.81x)
 2024-08-26 11:04:46 +02:00
Ramiro Polla
88a563ad18 swscale: export ff_copyPlane so it may be used by simd code 2024-08-26 11:04:46 +02:00
Ramiro Polla
4eb5594295 swscale: add nv24/nv42 to yuv420p unscaled converter 2024-08-26 11:04:46 +02:00
Martin Storsjö
cfe0a36352 libswscale: aarch64: Fix the indentation of some macro invocations 
Signed-off-by: Martin Storsjö <martin@martin.st>
 2024-08-22 14:40:30 +03:00
Martin Storsjö
507c2a5774 libswscale: arm: Don't assume aligned output in yuv2rgb functions 
This fixes failures in recently added checkasm tests.

While the buffers in most cases are aligned, libswscale in general
can't assume the output to be aligned.

Signed-off-by: Martin Storsjö <martin@martin.st>
 2024-08-19 23:04:52 +03:00
Ramiro Polla
181cd260db swscale/aarch64/yuv2rgb: add neon yuv42{0,2}p -> gbrp unscaled colorspace converters 
checkasm --bench on a Raspberry Pi 5 Model B Rev 1.0:
yuv420p_gbrp_128_c: 1243.0
yuv420p_gbrp_128_neon: 453.5
yuv420p_gbrp_1920_c: 18165.5
yuv420p_gbrp_1920_neon: 6700.0
yuv422p_gbrp_128_c: 1463.5
yuv422p_gbrp_128_neon: 471.5
yuv422p_gbrp_1920_c: 21343.7
yuv422p_gbrp_1920_neon: 6743.5
 2024-08-18 22:26:17 +02:00
Ramiro Polla
8744764a4c swscale/x86/yuv2rgb: add ssse3 yuv42{0,2}p -> gbrp unscaled colorspace converters 
Note: this implementation is limited to x86_64 due to general purpose
      register pressure.

checkasm --bench on an Intel(R) Core(TM) i5-5300U CPU @ 2.30GHz:
yuv420p_gbrp_8_c: 118.5
yuv420p_gbrp_8_ssse3: 93.3
yuv420p_gbrp_128_c: 1068.3
yuv420p_gbrp_128_ssse3: 319.3
yuv420p_gbrp_1080_c: 8841.8
yuv420p_gbrp_1080_ssse3: 2211.8
yuv420p_gbrp_1920_c: 15903.8
yuv420p_gbrp_1920_ssse3: 3814.3
yuv422p_gbrp_8_c: 144.8
yuv422p_gbrp_8_ssse3: 93.8
yuv422p_gbrp_128_c: 1395.8
yuv422p_gbrp_128_ssse3: 313.0
yuv422p_gbrp_1080_c: 11551.5
yuv422p_gbrp_1080_ssse3: 2240.8
yuv422p_gbrp_1920_c: 20585.3
yuv422p_gbrp_1920_ssse3: 5249.5
yuva420p_gbrp_8_c: 117.5
yuva420p_gbrp_8_ssse3: 92.0
yuva420p_gbrp_128_c: 1593.0
yuva420p_gbrp_128_ssse3: 319.3
yuva420p_gbrp_1080_c: 8694.5
yuva420p_gbrp_1080_ssse3: 2186.0
yuva420p_gbrp_1920_c: 15946.5
yuva420p_gbrp_1920_ssse3: 3805.3
 2024-08-18 22:26:14 +02:00
Ramiro Polla
4545205a26 swscale/yuv2rgb: add yuv42{0,2}p -> gbrp unscaled colorspace converters 2024-08-18 22:26:11 +02:00
Ramiro Polla
af5adf57e3 swscale/yuv2rgb: prepare YUV2RGBFUNC macro for multi-planar rgb 
This will be used in the upcoming yuv42{0,2}p -> gbrp unscaled
colorspace converters.

There is no difference in performance.
 2024-08-18 22:26:08 +02:00
Ramiro Polla
24063e7827 swscale/yuv2rgb: prepare LOADCHROMA/PUTFUNC macros for multi-planar rgb 
This will be used in the upcoming yuv42{0,2}p -> gbrp unscaled
colorspace converters.

There is no difference in performance.
 2024-08-18 22:26:05 +02:00
Niklas Haas
6b40be941a swscale/options: relax src/dst_h/v_chr_pos value range 
When dealing with 4x subsampling ratios (log2 == 2), such as can arise
with 4:1:1 or 4:1:0, a value range of 512 is not enough to cover the
range of possible scenarios.

For example, bottom-sited chroma in 4:1:0 would require an offset of 768
(three luma rows). Simply double the limit to 1024. I don't see any
place in initFilter() that would experience overflow as a result of this
change, especially since get_local_pos() right-shifts it by the
subsampling ratio again.
 2024-08-16 11:43:37 +02:00
Niklas Haas
3e064f52eb swscale: document SWS_FULL_CHR_H_* flags 
Based on my best understanding of what they do, given the source code.
 2024-08-16 11:43:37 +02:00
James Almer
66592e8b10 swscale/output: don't leave the alpha channel undefined in vuyx and xv36le 
It's non-determistic, as shown by poisoning avfilter buffers instead of zeroing them.

Signed-off-by: James Almer <jamrial@gmail.com>
 2024-08-13 14:49:41 -03:00
Rémi Denis-Courmont
210877c5fd sws/riscv: depend on RVB and simplify accordingly 2024-08-05 21:16:26 +03:00
Rémi Denis-Courmont
bd0c3edb13 lavu/riscv: count bytes rather than words for bswap32 
This removes the dependency on Zba at essentially zero cost.
 2024-07-30 18:41:51 +03:00
Shiyou Yin
4713a5cc24
swscale: [loongarch] Fix checkasm-sw_yuv2rgb failure. 
Reviewed-by: 陈昊 <chenhao@loongson.cn>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-07-28 19:02:16 +02:00
Rémi Denis-Courmont
4f2472909e sws/riscv: add forward-edge CFI landing pads 2024-07-25 23:10:14 +03:00
Rémi Denis-Courmont
e91a8cc4de sws/riscv: require B or zba explicitly 2024-07-25 18:55:48 +03:00
Michael Niedermayer
bcab9789ef
swscale/output: Fix integer overflows in yuv2rgba64_X_c_template 
Fixes: signed integer overflow: -1082982400 + -1068681048 cannot be represented in type 'int'
Fixes: 69995/clusterfuzz-testcase-minimized-ffmpeg_SWS_fuzzer-6285740271534080

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-07-21 15:35:08 +02:00
Niklas Haas
4ec45aca36 swscale/utils: fix leak on threaded ctx init failure 
This count gets incremented after init succeeds, when it should be
incremented after *alloc* succeeds. Otherwise, we leak the context on
failure.

There are no negative consequences of incrementing for
allocated-but-not-initialized contexts, as the only functions that
reference it will, in the worst case, simply behave as if called on
allocated-but-not-initialized contexts, which is in line with expected
behavior when sws_init_context() fails.
 2024-07-14 13:48:59 +02:00
Sean McGovern
34b4ca8696
swscale: prevent undefined behaviour in the PUTRGBA macro 
For even small values of 'asrc[x]', shifting them by 24 bits or more
will cause arithmetic overflow and be caught by
GCC's undefined behaviour sanitizer.

Ensure the values do not overflow by up-casting the bracketed
expressions involving 'asrc' to uint32_t.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-07-10 18:10:10 +02:00
Ramiro Polla
ac6263945a swscale/x86/yuv2rgb: Detemplatize 
Every function in yuv2rgb_template.c is only compiled exactly
once, so detemplatize it.
 2024-07-10 12:25:32 +02:00
Ramiro Polla
4f7f9b1026 swscale: remove unconditional #define DITHER1XBPP 
This seems to have had an use in the past, but it is now defined
unconditionally.
 2024-07-10 12:25:03 +02:00
Michael Niedermayer
66b60bae68
swscale/swscale: Use ptrdiff_t for linesize computations 
This is unlikely to make a difference

Fixes: CID1591896 Unintentional integer overflow
Fixes: CID1591901 Unintentional integer overflow

Sponsored-by: Sovereign Tech Fund
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-07-07 23:36:30 +02:00
Zhao Zhili
4d90a76986 swscale/aarch64: Add argb/abgr to yuv 
Test on Apple M1 with kperf:
				: -O3		: -O3 -fno-vectorize
abgr_to_uv_8_c			: 19.4		: 26.1
abgr_to_uv_8_neon		: 29.9		: 51.1
abgr_to_uv_128_c		: 146.4		: 558.9
abgr_to_uv_128_neon		: 85.1		: 83.4
abgr_to_uv_1080_c		: 1162.6	: 4786.4
abgr_to_uv_1080_neon		: 819.6		: 826.6
abgr_to_uv_1920_c		: 2063.6	: 8492.1
abgr_to_uv_1920_neon		: 1435.1	: 1447.1
abgr_to_uv_half_8_c		: 16.4		: 11.4
abgr_to_uv_half_8_neon		: 35.6		: 20.4
abgr_to_uv_half_128_c		: 108.6		: 359.4
abgr_to_uv_half_128_neon	: 75.4		: 42.6
abgr_to_uv_half_1080_c		: 883.4		: 2885.6
abgr_to_uv_half_1080_neon	: 460.6		: 481.1
abgr_to_uv_half_1920_c		: 1553.6	: 5106.9
abgr_to_uv_half_1920_neon	: 817.6		: 820.4
abgr_to_y_8_c			: 6.1		: 26.4
abgr_to_y_8_neon		: 40.6		: 6.4
abgr_to_y_128_c			: 99.9		: 390.1
abgr_to_y_128_neon		: 67.4		: 55.9
abgr_to_y_1080_c		: 735.9		: 3170.4
abgr_to_y_1080_neon		: 534.6		: 536.6
abgr_to_y_1920_c		: 1279.4	: 6016.4
abgr_to_y_1920_neon		: 932.6		: 927.6

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
 2024-07-05 16:32:31 +08:00
Zhao Zhili
52422133ae swscale/aarch64: Add bgra/rgba to yuv 
Test on Apple M1 with kperf
				: -O3		: -O3 -fno-vectorize
bgra_to_uv_8_c			: 13.4		: 27.5
bgra_to_uv_8_neon		: 37.4		: 41.7
bgra_to_uv_128_c		: 155.9		: 550.2
bgra_to_uv_128_neon		: 91.7		: 92.7
bgra_to_uv_1080_c		: 1173.2	: 4558.2
bgra_to_uv_1080_neon		: 822.7		: 809.5
bgra_to_uv_1920_c		: 2078.2	: 8115.2
bgra_to_uv_1920_neon		: 1437.7	: 1438.7
bgra_to_uv_half_8_c		: 17.9		: 14.2
bgra_to_uv_half_8_neon		: 37.4		: 10.5
bgra_to_uv_half_128_c		: 103.9		: 326.0
bgra_to_uv_half_128_neon	: 73.9		: 68.7
bgra_to_uv_half_1080_c		: 850.2		: 3732.0
bgra_to_uv_half_1080_neon	: 484.2		: 490.0
bgra_to_uv_half_1920_c		: 1479.2	: 4942.7
bgra_to_uv_half_1920_neon	: 824.2		: 824.7
bgra_to_y_8_c			: 8.2		: 29.5
bgra_to_y_8_neon		: 18.2		: 32.7
bgra_to_y_128_c			: 101.4		: 361.5
bgra_to_y_128_neon		: 74.9		: 73.7
bgra_to_y_1080_c		: 739.4		: 3018.0
bgra_to_y_1080_neon		: 613.4		: 544.2
bgra_to_y_1920_c		: 1298.7	: 5326.0
bgra_to_y_1920_neon		: 918.7		: 934.2

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
 2024-07-05 16:32:31 +08:00
Zhao Zhili
b8b71be07a swscale/aarch64: Add bgr24 to yuv 
Test on Apple M1 with kperf
				: -O3		: -O3 -fno-vectorize
bgr24_to_uv_8_c			: 28.5		: 52.5
bgr24_to_uv_8_neon		: 54.5		: 59.7
bgr24_to_uv_128_c		: 294.0		: 830.7
bgr24_to_uv_128_neon		: 99.7		: 112.0
bgr24_to_uv_1080_c		: 965.0		: 6624.0
bgr24_to_uv_1080_neon		: 751.5		: 754.7
bgr24_to_uv_1920_c		: 1693.2	: 11554.5
bgr24_to_uv_1920_neon		: 1292.5	: 1307.5
bgr24_to_uv_half_8_c		: 54.2		: 37.0
bgr24_to_uv_half_8_neon		: 27.2		: 22.5
bgr24_to_uv_half_128_c		: 127.2		: 392.5
bgr24_to_uv_half_128_neon	: 63.0		: 52.0
bgr24_to_uv_half_1080_c		: 880.2		: 3329.0
bgr24_to_uv_half_1080_neon	: 401.5		: 390.7
bgr24_to_uv_half_1920_c		: 1585.7	: 6390.7
bgr24_to_uv_half_1920_neon	: 694.7		: 698.7
bgr24_to_y_8_c			: 21.7		: 22.5
bgr24_to_y_8_neon		: 797.2		: 25.5
bgr24_to_y_128_c		: 88.0		: 280.5
bgr24_to_y_128_neon		: 63.7		: 55.0
bgr24_to_y_1080_c		: 616.7		: 2208.7
bgr24_to_y_1080_neon		: 900.0		: 452.0
bgr24_to_y_1920_c		: 1093.2	: 3894.7
bgr24_to_y_1920_neon		: 777.2		: 767.5

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
 2024-07-05 16:32:31 +08:00
Ramiro Polla
61e851381f swscale/yuv2rgb/x86: remove mmx/mmxext yuv2rgb functions 
These functions are either slower or barely faster than the C LUT
yuv2rgb code.
 2024-07-04 11:12:47 +02:00
Michael Niedermayer
c221c7422f
swscale/output: Avoid undefined overflow in yuv2rgb_write_full() 
Fixes: signed integer overflow: -140140 * 16525 cannot be represented in type 'int'
Fixes: 68859/clusterfuzz-testcase-minimized-ffmpeg_SWS_fuzzer-4516387130245120

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-06-26 20:49:36 +02:00
Michael Niedermayer
9e6c5b6e86
swscale/output: alpha can become negative after scaling, use multiply 
Fixes: left shift of negative value -3245
Fixes: 69047/clusterfuzz-testcase-minimized-ffmpeg_SWS_fuzzer-6571511551950848

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
 2024-06-26 20:49:36 +02:00
Ramiro Polla
e37a93031e swscale/yuv2rgb: reindent after previous commit 2024-06-24 13:35:22 +02:00
Ramiro Polla
0a08c64588 swscale/yuv2rgb: fix yuv422p input in C code 
The C code was silently ignoring the second chroma line on yuv422p
input.
 2024-06-24 13:34:53 +02:00
Ramiro Polla
fb8fae864f swscale/yuv2rgb: add macros to simplify code generation 2024-06-24 13:34:28 +02:00
Ramiro Polla
88a402df74 swscale/yuv2rgb: fix conversion for widths not aligned to 8 
The C code for some pixel formats (rgb555, rgb565, rgb444, and monob)
was not converting the last pixels on widths not aligned to 8.

NOTE: the last pixel for odd widths is still not converted for any of
      the pixel formats in the C code for yuv2rgb except for monob.
 2024-06-24 13:33:53 +02:00
Ramiro Polla
75f1a8e071 swscale/aarch64: add neon {lum,chr}ConvertRange 
chrRangeFromJpeg_8_c: 29.2
chrRangeFromJpeg_8_neon: 19.5
chrRangeFromJpeg_24_c: 80.5
chrRangeFromJpeg_24_neon: 34.0
chrRangeFromJpeg_128_c: 413.7
chrRangeFromJpeg_128_neon: 156.0
chrRangeFromJpeg_144_c: 471.0
chrRangeFromJpeg_144_neon: 174.2
chrRangeFromJpeg_256_c: 842.0
chrRangeFromJpeg_256_neon: 305.5
chrRangeFromJpeg_512_c: 1699.0
chrRangeFromJpeg_512_neon: 608.0
chrRangeToJpeg_8_c: 51.7
chrRangeToJpeg_8_neon: 22.7
chrRangeToJpeg_24_c: 149.7
chrRangeToJpeg_24_neon: 38.0
chrRangeToJpeg_128_c: 761.7
chrRangeToJpeg_128_neon: 176.7
chrRangeToJpeg_144_c: 866.2
chrRangeToJpeg_144_neon: 198.7
chrRangeToJpeg_256_c: 1516.5
chrRangeToJpeg_256_neon: 348.7
chrRangeToJpeg_512_c: 3067.2
chrRangeToJpeg_512_neon: 692.7
lumRangeFromJpeg_8_c: 24.0
lumRangeFromJpeg_8_neon: 17.0
lumRangeFromJpeg_24_c: 56.7
lumRangeFromJpeg_24_neon: 21.0
lumRangeFromJpeg_128_c: 294.5
lumRangeFromJpeg_128_neon: 76.7
lumRangeFromJpeg_144_c: 332.5
lumRangeFromJpeg_144_neon: 86.7
lumRangeFromJpeg_256_c: 586.0
lumRangeFromJpeg_256_neon: 152.2
lumRangeFromJpeg_512_c: 1190.0
lumRangeFromJpeg_512_neon: 298.0
lumRangeToJpeg_8_c: 31.7
lumRangeToJpeg_8_neon: 19.5
lumRangeToJpeg_24_c: 83.5
lumRangeToJpeg_24_neon: 24.2
lumRangeToJpeg_128_c: 440.5
lumRangeToJpeg_128_neon: 91.0
lumRangeToJpeg_144_c: 504.2
lumRangeToJpeg_144_neon: 101.0
lumRangeToJpeg_256_c: 879.7
lumRangeToJpeg_256_neon: 177.2
lumRangeToJpeg_512_c: 1794.2
lumRangeToJpeg_512_neon: 354.0
 2024-06-18 23:12:41 +02:00
James Almer
fcf72966a5 swscale/x86/range_convert: add missing AVX2 preprocessor wrapper 
Fixes compilation with old yasm.

Signed-off-by: James Almer <jamrial@gmail.com>
 2024-06-16 10:09:38 -03:00
James Almer
8a4c9d6bd3 swscale/x86/range_convert: reduce amount of xmm regs clobbered in luma functions 
Signed-off-by: James Almer <jamrial@gmail.com>
 2024-06-15 21:02:06 -03:00
Ramiro Polla
f6859cade3 swscale/x86: add sse2 and avx2 {lum,chr}ConvertRange 
chrRangeFromJpeg_8_c: 22.3
chrRangeFromJpeg_8_sse2: 13.3
chrRangeFromJpeg_8_avx2: 13.3
chrRangeFromJpeg_24_c: 72.8
chrRangeFromJpeg_24_sse2: 22.3
chrRangeFromJpeg_24_avx2: 17.5
chrRangeFromJpeg_128_c: 345.5
chrRangeFromJpeg_128_sse2: 106.0
chrRangeFromJpeg_128_avx2: 57.8
chrRangeFromJpeg_144_c: 380.5
chrRangeFromJpeg_144_sse2: 118.5
chrRangeFromJpeg_144_avx2: 62.3
chrRangeFromJpeg_256_c: 646.3
chrRangeFromJpeg_256_sse2: 218.8
chrRangeFromJpeg_256_avx2: 109.0
chrRangeFromJpeg_512_c: 1461.5
chrRangeFromJpeg_512_sse2: 426.5
chrRangeFromJpeg_512_avx2: 211.5
chrRangeToJpeg_8_c: 37.8
chrRangeToJpeg_8_sse2: 10.5
chrRangeToJpeg_8_avx2: 14.0
chrRangeToJpeg_24_c: 114.3
chrRangeToJpeg_24_sse2: 23.5
chrRangeToJpeg_24_avx2: 16.3
chrRangeToJpeg_128_c: 633.5
chrRangeToJpeg_128_sse2: 107.5
chrRangeToJpeg_128_avx2: 55.0
chrRangeToJpeg_144_c: 758.3
chrRangeToJpeg_144_sse2: 132.0
chrRangeToJpeg_144_avx2: 64.5
chrRangeToJpeg_256_c: 1345.0
chrRangeToJpeg_256_sse2: 218.0
chrRangeToJpeg_256_avx2: 105.3
chrRangeToJpeg_512_c: 2524.0
chrRangeToJpeg_512_sse2: 417.0
chrRangeToJpeg_512_avx2: 218.8
lumRangeFromJpeg_8_c: 11.8
lumRangeFromJpeg_8_sse2: 11.0
lumRangeFromJpeg_8_avx2: 10.3
lumRangeFromJpeg_24_c: 38.5
lumRangeFromJpeg_24_sse2: 15.5
lumRangeFromJpeg_24_avx2: 12.5
lumRangeFromJpeg_128_c: 232.3
lumRangeFromJpeg_128_sse2: 60.0
lumRangeFromJpeg_128_avx2: 26.8
lumRangeFromJpeg_144_c: 259.5
lumRangeFromJpeg_144_sse2: 65.3
lumRangeFromJpeg_144_avx2: 29.0
lumRangeFromJpeg_256_c: 464.5
lumRangeFromJpeg_256_sse2: 107.5
lumRangeFromJpeg_256_avx2: 54.0
lumRangeFromJpeg_512_c: 897.5
lumRangeFromJpeg_512_sse2: 224.5
lumRangeFromJpeg_512_avx2: 109.8
lumRangeToJpeg_8_c: 17.8
lumRangeToJpeg_8_sse2: 11.0
lumRangeToJpeg_8_avx2: 11.8
lumRangeToJpeg_24_c: 56.3
lumRangeToJpeg_24_sse2: 11.0
lumRangeToJpeg_24_avx2: 12.5
lumRangeToJpeg_128_c: 333.8
lumRangeToJpeg_128_sse2: 53.3
lumRangeToJpeg_128_avx2: 26.5
lumRangeToJpeg_144_c: 375.5
lumRangeToJpeg_144_sse2: 60.8
lumRangeToJpeg_144_avx2: 29.0
lumRangeToJpeg_256_c: 652.0
lumRangeToJpeg_256_sse2: 109.5
lumRangeToJpeg_256_avx2: 53.5
lumRangeToJpeg_512_c: 1284.3
lumRangeToJpeg_512_sse2: 218.0
lumRangeToJpeg_512_avx2: 108.3
 2024-06-16 00:35:51 +02:00
Rémi Denis-Courmont
378d1b06c3 riscv: probe for Zbb extension at load time 
Due to hysterical raisins, most RISC-V Linux distributions target a
RV64GC baseline excluding the Bit-manipulation ISA extensions, most
notably:
- Zba: address generation extension and
- Zbb: basic bit manipulation extension.
Most CPUs that would make sense to run FFmpeg on support Zba and Zbb
(including the current FATE runner), so it makes sense to optimise for
them. In fact a large chunk of existing assembler optimisations relies
on Zba and/or Zbb.

Since we cannot patch shared library code, the next best thing is to
carry a flag initialised at load-time and check it on need basis.
This results in 3 instructions overhead on isolated use, e.g.:
1:  AUIPC rd, %pcrel_hi(ff_rv_zbb_supported)
    LBU   rd, %pcrel_lo(1b)(rd)
    BEQZ  rd, non_Zbb_fallback_code
    // Zbb code here

The C compiler will typically load the flag ahead of time to reducing
latency, and can also keep it around if Zbb is used multiple times in a
single optimisation scope. For this to work, the flag symbol must be
hidden; otherwise the optimisation degrades with a GOT look-up to
support interposition:
1:  AUIPC rd, GOT_OFFSET_HI
    LD    rd, GOT_OFFSET_LO(rd)
    LBU   rd, (rd)
    BEQZ  rd, non_Zbb_fallback_code
    // Zbb code here

This patch adds code to provision the flag in libraries using bit
manipulation functions from libavutil: byte-swap, bit-weight and
counting leading or trailing zeroes.
 2024-06-11 20:12:37 +03:00
Rémi Denis-Courmont
417957ec5e sws/range_convert: R-V V to/from JPEG 
C908   X60
chrRangeFromJpeg_8_c:          2.7    2.5
chrRangeFromJpeg_8_rvv_i32:    1.7    1.5
chrRangeFromJpeg_24_c:         7.5    6.7
chrRangeFromJpeg_24_rvv_i32:   1.7    1.5
chrRangeFromJpeg_128_c:       55.2   34.7
chrRangeFromJpeg_128_rvv_i32:  6.5    3.0
chrRangeFromJpeg_144_c:       44.0   39.2
chrRangeFromJpeg_144_rvv_i32:  7.7    4.5
chrRangeFromJpeg_256_c:       78.2   69.5
chrRangeFromJpeg_256_rvv_i32: 12.2    6.0
chrRangeFromJpeg_512_c:      172.2  138.5
chrRangeFromJpeg_512_rvv_i32: 24.5   11.7
chrRangeToJpeg_8_c:            4.7    4.2
chrRangeToJpeg_8_rvv_i32:      2.0    1.7
chrRangeToJpeg_24_c:          13.7   12.2
chrRangeToJpeg_24_rvv_i32:     2.0    1.5
chrRangeToJpeg_128_c:         72.0   63.7
chrRangeToJpeg_128_rvv_i32:    6.7    3.2
chrRangeToJpeg_144_c:         80.7   71.7
chrRangeToJpeg_144_rvv_i32:    8.5    4.7
chrRangeToJpeg_256_c:        143.2  127.2
chrRangeToJpeg_256_rvv_i32:   13.5    6.5
chrRangeToJpeg_512_c:        285.7  253.7
chrRangeToJpeg_512_rvv_i32:   27.0   13.0
lumRangeFromJpeg_8_c:          1.7    1.5
lumRangeFromJpeg_8_rvv_i32:    1.2    1.0
lumRangeFromJpeg_24_c:         4.2    3.7
lumRangeFromJpeg_24_rvv_i32:   1.2    1.0
lumRangeFromJpeg_128_c:       21.7   19.2
lumRangeFromJpeg_128_rvv_i32:  3.7    1.7
lumRangeFromJpeg_144_c:       24.7   22.0
lumRangeFromJpeg_144_rvv_i32:  4.7    2.7
lumRangeFromJpeg_256_c:       43.7   39.0
lumRangeFromJpeg_256_rvv_i32:  7.5    3.2
lumRangeFromJpeg_512_c:       87.0   77.2
lumRangeFromJpeg_512_rvv_i32: 14.5    6.7
lumRangeToJpeg_8_c:            2.7    2.2
lumRangeToJpeg_8_rvv_i32:      1.0    1.0
lumRangeToJpeg_24_c:           7.2    6.5
lumRangeToJpeg_24_rvv_i32:     1.2    1.0
lumRangeToJpeg_128_c:         37.7   33.7
lumRangeToJpeg_128_rvv_i32:    3.7    2.0
lumRangeToJpeg_144_c:         42.5   37.7
lumRangeToJpeg_144_rvv_i32:    4.7    2.7
lumRangeToJpeg_256_c:         75.0   66.7
lumRangeToJpeg_256_rvv_i32:    7.5    3.5
lumRangeToJpeg_512_c:        149.5  133.0
lumRangeToJpeg_512_rvv_i32:   14.7    7.0
 2024-06-10 22:48:52 +03:00
Zhao Zhili
9dac8495b0 swscale/aarch64: Add rgb24 to yuv implementation 
Test on Apple M1:

rgb24_to_uv_8_c: 0.0
rgb24_to_uv_8_neon: 0.2
rgb24_to_uv_128_c: 1.0
rgb24_to_uv_128_neon: 0.5
rgb24_to_uv_1080_c: 7.0
rgb24_to_uv_1080_neon: 5.7
rgb24_to_uv_1920_c: 12.5
rgb24_to_uv_1920_neon: 9.5
rgb24_to_uv_half_8_c: 0.2
rgb24_to_uv_half_8_neon: 0.2
rgb24_to_uv_half_128_c: 1.0
rgb24_to_uv_half_128_neon: 0.5
rgb24_to_uv_half_1080_c: 6.2
rgb24_to_uv_half_1080_neon: 3.0
rgb24_to_uv_half_1920_c: 11.2
rgb24_to_uv_half_1920_neon: 5.2
rgb24_to_y_8_c: 0.2
rgb24_to_y_8_neon: 0.0
rgb24_to_y_128_c: 0.5
rgb24_to_y_128_neon: 0.5
rgb24_to_y_1080_c: 4.7
rgb24_to_y_1080_neon: 3.2
rgb24_to_y_1920_c: 8.0
rgb24_to_y_1920_neon: 5.7

On Pixel 6:

rgb24_to_uv_8_c: 30.7
rgb24_to_uv_8_neon: 56.9
rgb24_to_uv_128_c: 213.9
rgb24_to_uv_128_neon: 173.2
rgb24_to_uv_1080_c: 1649.9
rgb24_to_uv_1080_neon: 1424.4
rgb24_to_uv_1920_c: 2907.9
rgb24_to_uv_1920_neon: 2480.7
rgb24_to_uv_half_8_c: 36.2
rgb24_to_uv_half_8_neon: 33.4
rgb24_to_uv_half_128_c: 167.9
rgb24_to_uv_half_128_neon: 99.4
rgb24_to_uv_half_1080_c: 1293.9
rgb24_to_uv_half_1080_neon: 778.7
rgb24_to_uv_half_1920_c: 2292.7
rgb24_to_uv_half_1920_neon: 1328.7
rgb24_to_y_8_c: 19.7
rgb24_to_y_8_neon: 27.7
rgb24_to_y_128_c: 129.9
rgb24_to_y_128_neon: 96.7
rgb24_to_y_1080_c: 995.4
rgb24_to_y_1080_neon: 767.7
rgb24_to_y_1920_c: 1747.4
rgb24_to_y_1920_neon: 1337.2

Note both tests use clang as compiler, which has vectorization
enabled by default with -O3.

Reviewed-by: Rémi Denis-Courmont <remi@remlab.net>
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
 2024-06-11 01:12:09 +08:00
James Almer
17c3cc5bb6 swscale/x86/rgb_2_rgb: add missing wrap to ff_uyvytoyuv422_avx2 
Fixes old yasm.

Signed-off-by: James Almer <jamrial@gmail.com>
 2024-06-09 16:04:36 -03:00